Note—Estimating Geographic Customer Densities Using Kernel Density Estimation

This paper shows how kernel density estimation may be used to estimate flexibly the geographic distribution of customers in a market. In addition it shows how a density-based product positioning methodology may be applied to site selection, using the estimated geographic customer density to help locate a new or (relocated) store or distribution center. This application provides a conceptual basis for more complicated site selection and spatial demand models which might involve several predictor variables.